Cost Comparison of B-1B Non-Mission-Capable Drivers Using Finite Source Queueing with Spares.

摘要

Maintenance costs and maintenance man-hours have increased dramatically in the last couple of decades in relation to flying hours. These increased costs, combined with shrinking budgets, force the Air Force to streamline maintenance processes and be selective concerning which maintenance processes should receive additional funding. There are many drivers rendering an aircraft non-mission-capable (NMC). This research provides a method to compare the cost associated with any NMC driver with other NMC drivers in order to determine where limited resources are best allocated towards the goal of finding more efficient solutions that also result in reduced cost. The cost model includes lost flying time, maintenance, and parts making it more comprehensive than current methods. Evaluation of the cost function requires estimating both number of aircraft out of service and time out of service given the behavior of the maintenance system. This is compounded by the fact that there are a small number of aircraft in a flying wing. These aircraft are split between missions and preventative maintenance. Furthermore, due to the increased age of the fleet, the aircraft prepped for missions aren't always mission capable requiring extra aircraft be prepped and ready to step into the lineup making large-number approximations unusable. Instead, a finite source queueing model including spares is incorporated resulting in simple-to-use calculations requiring no special computational resources or training. In fact, as the detailed sensitivity analyses provided in this research demonstrate, the comparison of multiple NMC drivers using the provided cost function is fairly simple provided a reliable estimate of the associated data. The specific application of the analysis undertaken with this cost/queue formulation is the B-1B bomber. Complete maintenance data from the 28th Maintenance Operations Squadron over 5 years is analyzed to define the parameters of the model and validate its results. R.